To determine if obese, hyperglycemic, non-ketotic teenagers are as insulin-resistant as adults with type 2 diabetes and if any of these patients have evidence of autoimmunity characteristic of type 1 diabetes. To determine the long-term efficiacy of metformin in this subset of diabetic patients and whether peripheral amylin/insulin ration distinguishes these patients and predicts their response to metformin.